Editor’s Note: I wrote this story and then saw the news he had been sent down.
As a general rule, people mistrust new things and nowhere is this more evident than in baseball statistics. I am among a generation of people who grew up getting my baseball numbers primarily from the newspaper, baseball cards and television broadcasts. These three sources told me that AVG, HR and RBIs were the important hitting statistics, while W-L, ERA and Ks were the pitching ones upon which to focus.
And to a varying degree, those are all important statistics. It’s just that they don’t tell us as much about winning baseball as we thought when we were kids.
When new metrics are introduced, they are held to a standard that is almost impossible to meet. People focus on the problems of the new stats and then go scurrying back to the tried and true Triple Crown numbers instead, often times failing to understand where these traditional stats come up short.
One of the keys to new statistics is to understand how and when to utilize them. Just like you would not use HR to judge how fast a player is, many new stats have a specific purpose in mind. Today I want to talk about a new stat called WPA. This is far from my favorite stat but it does have its uses.
WPA stands for Win Probability Added. It sounds complex but it is really an intuitive metric. WPA measures how a plate appearance (or series of plate appearances if you’re looking at more than one) changes the percent chance of a team winning the game. This stat is based on a bunch of different things, including, but not limited to, the score, the inning, number of outs and runners on base.
For batters a walk in the third inning of a game where you are winning 10-0 likely will have a WPA near zero. A grand slam home run in the bottom of the ninth when your team trails by three runs will have a WPA near 1.000
And as we’ll see in a minute, failing in big situations will give you a negative WPA.
Some people have compared WPA to telling a story of the game. Others think of it as a measurement of clutch ability. It certainly has elements of both of these things. What it does do is give a nice picture of what really happened. However, it has no predictive ability and it measures opportunity at least as much (if not more so) than skill.
Now I want to use WPA to talk about Chin-lung Hu.
In last night’s game, Hu was brought on in the ninth inning of a tie game with two outs and the bases loaded. He grounded into a fielder’s choice. His WPA for this plate appearance was -.151 or in other words, the Mets had a 65 percent chance to win the game when Hu was at the plate and after he made out it dropped to 50 percent.
It was one of four plays in the game that dropped the Mets’ chances of winning by 10 percent or more. The biggest play was Justin Turner grounding into the double play in the 10th inning (-.196). The other two big plays were Jose Reyes striking out to end the game with the tying run on third base (-.148) and Jason Pridie striking out with a runner on second base in the ninth (-.112).
Here are Hu’s WPA for each at-bat he’s had in the month of May:
| 5/1 | K | -.043 |
| 5/3 | F8 | -.056 |
| 5/5 | K | -.019 |
| 5/5 | K | -.018 |
| 5/5 | P5 | -.003 |
| 5/5 | BB | .024 |
| 5/8 | K | -.041 |
| 5/14 | P4 | -.012 |
| 5/16 | FC | -.151 |
In his last nine plate appearances, Hu has a -0.319 WPA. He’s not just making outs, he’s making outs in important parts of the ballgame. For the year, Hu has a -0.41 WPA. If we break his PA down into positive contributions and negative contributions, Hu’s PA where he’s pushed the team forward equal 0.15 while the times he’s hurt the team equal -0.56
For a comparison, Brad Emaus had an overall WPA of -0.44 (+0.32 and -0.77) and he was sent packing.
There’s very little recommending that Hu be kept on the major league roster. He’s barely needed as a backup shortstop and the Mets have several players (Harris, Murphy, Turner) capable of playing second base. It’s more important that the Mets get a bat on the bench who can occasionally deliver a base hit.
Hu’s overall record indicates a player not likely to hit major league pitching. In 214 PA in the majors, Hu has a .176/.225/.259 slash line which is horrible. Because the Mets are carrying seven relievers, they have only five bench players. With David Wright unable to go last night, they were down to four. Terry Collins seemingly had no option when he used Hu in the ninth inning and the results were predictably bad.
They were made to look even worse when Jonathon Niese came on as a pinch-hitter in the 11th inning and delivered a triple. Hu was the on-deck hitter when Reyes struck out to end the game. It’s a reasonable question to ask if Collins should have pinch-hit a pitcher instead of let Hu come up with the game on the line again if Reyes had extended the game.
Hu’s WPA shows how much he is hurting the team right now. His career numbers indicate a player not likely to hit major league pitching in the future. His role as backup infielder is so much less important right now than his role as pinch-hitter. And he’s been a miserable failure as a pinch-hitter.
It’s time, past time really, to end the Hu experiment.